Included Examples
| Cookbook | Graph Type | Task |
|---|---|---|
| Banking77 Classification | Policy | Intent classification (77 classes) |
| Crafter Verifier | Verifier | Custom judge for RL agent traces |
Policy vs Verifier
Policy graphs solve tasks - they take inputs and produce outputs:- Banking77:
{customer_message}→{intent}
- Crafter:
{trace}→{score, reasoning}
Each Cookbook Includes
- An
ADASTaskSetJSON dataset - Runnable training code
- Inference examples
- Notes on judge configuration
Getting Started
- Pick the cookbook matching your use case
- Copy the dataset format
- Adapt to your domain
- Train with ADAS